Low-rank adaptation lora - Open Access .click

Results 81 to 90 of about 9,649,421 (248)

A Review and Prospect of Appearance Defect Detection Methods for High Voltage Equipment Based on Vision Large Models

High Voltage, EarlyView.
ABSTRACT High voltage equipment, as a core component of power systems, plays an indispensable role in ensuring the reliability of power supply through its safe and stable operation. Traditional visual defect detection for high voltage equipment, however, often relies on manual inspection and experience‐based judgement, which struggles to meet the ...
Zhenbing Zhao +7 more
wiley +1 more source

LoRA-Adv: Boosting Text Classification in Large Language Models Through Adversarial Low-Rank Adaptations

IEEE Access
Low-rank adaptation (LoRA), a paradigm bridging the gap between large language models and fine-tuning, has demonstrated effectiveness across various natural language processing tasks.
Hong Ye +4 more
doaj +1 more source

A validated 3D point‐cloud analysis workflow for the creation of seagrass canopy height models

Methods in Ecology and Evolution, EarlyView.
Abstract Seagrasses support highly productive ecosystems in both temperate and tropical waters and provide essential ecosystem services. Despite their importance, the vertical structure of seagrasses (i.e. the canopy) remains poorly characterised at the microscale, even though it is a strong predictor of ecological processes that influence both ...
Daniele Ventura
wiley +1 more source

MiLO-LoRA: Adaptive Low-rank Adaptation Allocation on Large Language Models

Low-rank adaptation (LoRA) is a representative method in the field of parameter-efficient fine-tuning (PEFT), and is key to democratizating the modern large language models (LLMs). The vanilla LoRA is implemented with uniform ranks, and the recent literature have found that properly allocating LoRA ranks on the LLM backbones results in performance ...
Yuan, Zhe +4 more
openaire +2 more sources

Future of edge AI in biodiversity monitoring

Methods in Ecology and Evolution, EarlyView.
Abstract Many ecological decisions are slowed by the gap between collecting and analysing biodiversity information. Edge computing moves data processing closer to the sensor, with edge artificial intelligence (AI) using AI models for data processing and selective data transfer.
Aude Vuilliomenet, Kate E. Jones, Duncan Wilson +2 more
wiley +1 more source

Fine-Tuning Strategies for Sentiment Analysis in the Algerian Dialect: A Comparative Study on DziriBERT

ITEGAM-JETIA
The aim of this work is to explore sentiment analysis in the Algerian dialect through the adaptation of pre-trained linguistic models. We focus on DziriBERT, a model specifically developed for the Algerian dialect and derived from the BERT ...
Salima Brachemi-Meftah, Fatiha Barigou
doaj +1 more source

Online-LoRA: Task-Free Online Continual Learning via Low Rank Adaptation

2025 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV)
Catastrophic forgetting is a significant challenge in online continual learning (OCL), especially for non-stationary data streams that do not have well-defined task boundaries. This challenge is exacerbated by the memory constraints and privacy concerns inherent in rehearsal buffers. To tackle catastrophic forgetting, in this paper, we introduce Online-
Xiwen Wei, Guihong Li, Radu Marculescu
openaire +4 more sources

Tipping the Scales Using Cryopreservation: Modelling the Genetic Value and Management Potential of Integrated Cryopreservation for Threatened Ex‐Situ Pupfish Populations

Animal Conservation, EarlyView.
This graphical abstract depicts the number of additional fish that would be required to achieve 90% genetic diversity for endangered pupfish species housed in EAZA collections, with or without the use of cryopreserved material. ABSTRACT Zoos and aquaria must retain genetic diversity for the species they house if the populations are to remain viable ...
James Edward Brereton +5 more
wiley +1 more source

LoTR: Low Tensor Rank Weight Adaptation [PDF]

In this paper we generalize and extend an idea of low-rank adaptation (LoRA) of large language models (LLMs) based on Transformer architecture. Widely used LoRA-like methods of fine-tuning LLMs are based on matrix factorization of gradient update.
Cherniuk, Daria +4 more
core +1 more source

Effect of intensive care unit sedation strategy on patient‐reported experience: secondary mediation analysis of the A2B randomised controlled trial

Anaesthesia, EarlyView.
Summary Introduction The optimal ICU sedation strategy to improve patient‐reported experience is unknown. In a planned secondary analysis of the A2B randomised trial, we evaluated the effect of sedation strategy (dexmedetomidine, clonidine or propofol) on recalled patient experience at 90 days post‐randomisation.
John A. Masterson +657 more
wiley +1 more source

fos: computer and information sciences
machine learning cs.lg
low-rank adaptation

artificial intelligence cs.ai
parameter-efficient fine-tuning
computer science - machine learning

machine learning
computation and language cs.cl
lora